Fitting a Function with trainlm to find function parameters?

I would like to fit a curve with the Levenberg-Marquardt (LM) Algorithm, and I was wondering if it is possible to use the trainlm algorithm?

For example, I would like to fit a set of data points to with:

f(x) = a*sin(x) + b*cos(x).

I would like to estimate a and b using the LM algorithm in Matlab. Is this possible simply with the trainlm algorithm?

I have found a variety of sources, like: Curve Fitting via Optimization, but I am unable to Open the Script to view the source code.

Answer 1

[Matt J]

Why not just use the Levenberg-Marquardt option of LSQCURVEFIT? Also, I hope you know that your example, a linear estimation problem, is far too trivial to bother with Levenberg-Marquardt. You could do it analytically with MLDIVIDE.

Answer 2

[Greg Heath]

Think of a neural net as a mapping between I-dimensional "I"nputs and O-dimensional "O"utputs achieved by using N pairs of I-dimensional input examples and corresponding O-dimensional output target examples.

Your problem does not fit into that category.

Hope this helps.

Thank you for formally accepting my answer

Greg